摘要
针对黏菌算法收敛精度较低、易陷入局部最优和收敛速度较慢的问题,提出了一种动态非线性参数的反向学习黏菌算法.利用反向学习策略丰富种群多样性,获得更好的种群作为初始种群,提高算法优化性能和收敛速度;提出动态非线性递减策略,动态调整黏菌搜索区域;平衡算法全局探索和局部开发能力,提高算法跳出局部最优解的能力,提高收敛精度.不同算法之间的实验对比是使用几个基准测试函数进行的,结果表明,改进算法具有更强的寻优特性和更快的收敛速度,并且收敛精度得到了不同程度的提高.最后通过2个工程设计问题验证了改进算法在实际应用问题中的可靠性和有效性.
In response to the low convergence accuracy,propensity for local optima,and slow convergence speed of the slime mold algorithm,a dynamic nonlinear parameter opposition-based learning slime mold algorithm is proposed.By using a opposition-based learning strategy to enrich population diversity and obtain a better initial population,the algorithm′s optimization performance and convergence speed are improved.A dynamic nonlinear decreasing strategy is introduced to dynamically adjust the slime mold search area,to coordinate global exploration and local development to enhance the algorithm′s ability to avoid local optima and to improve convergence accuracy.Experimental comparisons between different algorithms are conducted by using several benchmark test functions.The results show that the improved algorithm has stronger optimization characteristics and faster convergence speed,with varying degrees of improvement in convergence accuracy.Finally,the reliability and effectiveness of the improved algorithm in practical application problems are validated through two engineering design problems.
作者
饶爽
曲良东
梅雨琳
RAO Shuang;QU Liangdong;MEI Yulin(College of Electronic Information,Guangxi Minzu University,Nanning 530006,China;College of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,China)
出处
《成都大学学报（自然科学版）》
2024年第4期371-378,共8页
Journal of Chengdu University（Natural Science Edition）
基金
广西自然科学基金项目(2023GXNSFBA026019、2019GXNSFAA185033)
广西科技基地和人才专项(GUIKE AD18126010、GUIKEAD22080021)。
关键词
反向学习
动态非线性递减
黏菌算法
算法优化
opposition-based learning
dynamic nonlinear decreasing
slime mold algorithm
algorithm optimization
作者简介
饶爽(1999-),男,硕士研究生,从事智能优化算法研究.E-mail:1064882740@qq.com;通信作者:曲良东(1976-),男,硕士,副教授,从事智能优化算法、机器学习和粗糙集研究.E-mail:quliangdong@126.com。
